which I think should be our choice, there are good examples these days. Servers can even register as service providers. XMPP is really not designed for this application.
On Wed, Jun 12, 2013 at 7:18 PM, Yuri Z <vega...@gmail.com> wrote: > But without XMPP you would need to define your own discovery protocol. > > > On Wed, Jun 12, 2013 at 4:47 PM, Michael MacFadden < > michael.macfad...@gmail.com> wrote: > > > Correct, it's just another configuration point. > > > > On 6/12/13 2:44 PM, "Dave" <w...@glark.co.uk> wrote: > > > > >On 12/06/13 12:43, Michael MacFadden wrote: > > >> So, most people would not argue [against] XMPP as an established chat > > >>protocol. > > >> However, wave is not chat. > > > > > >Indeed. Xmpp is widely deployed for chat, but not widely deployed for > > >wave. Wave traffic is more noisy than simple chat, but the S2S > > >connections for a busy xmpp server are also pretty noisy. > > > > > >Apologies if my earlier comment appeared to imply otherwise. > > > > > >We're using xmpp in a unusual way, and it's of questionable benefit to > > >wave. But the only thing I've seen that we can truly blaim on xmpp is > > >that the server/component configuration is unnecessarily painful. > > > > > > > > >Dave > > > > > > >
